[email protected] (Jon Perryman) writes:
> 2. It is the only tool where we can easilyt segregate interactive
> versus long running programs. This allows WLM give more resources to
> interactive users because they are personally waiting. Sysprog's
> encourage it's use by setting WLM such that a user get's less than
> batch priority when they use to many resources.

re:
http://www.garlic.com/~lynn/2013n.html#16 z/OS is antique WAS: Aging Sysprogs = 
Aging Farmers
http://www.garlic.com/~lynn/2013n.html#17 z/OS is antique WAS: Aging Sysprogs = 
Aging Farmers
http://www.garlic.com/~lynn/2013n.html#18 z/OS is antique WAS: Aging Sysprogs = 
Aging Farmers
http://www.garlic.com/~lynn/2013n.html#19 z/OS is antique WAS: Aging Sysprogs = 
Aging Farmers

as an undergraduate in the 60s, I did dynamic adaptive resource manager
for cp67 ... which was picked up and released as part of the product.  A
default policy was "fair share" resources ... nobody got more resources
than anybody else ... regardless of interactive or background/batch
characteristics ... default policy gave interactive more timely
resources ... but not more resources. some past posts
http://www.garlic.com/~lynn/subtopic.html#fairshare

in the simplification morph from cp67 to vm370, the dynamic adaptive
code was dropped ... however customers would continue to advocate in
share to bring it back.

I went to the science center ... some past posts
http://www.garlic.com/~lynn/subtopic.html#545tech

and continued to work on cp67 and then did port to vm370
http://www.garlic.com/~lynn/2006v.html#email731212
http://www.garlic.com/~lynn/2006w.html#email750102
http://www.garlic.com/~lynn/2006w.html#email750430

one of my hobbies in IBM was producing, distributing & supporting
production systems for large number of internal datacenters
... including the internal IBM world-wide sales&marketing HONE systems
... some past posts
http://www.garlic.com/~lynn/subtopic.html#hone

this was also during the Future System period ... when 370 efforts were
being shutdown ... but I continued with 360/370 and not exactly career
enhancing ... critized what they were doing ... some past posts
http://www.garlic.com/~lynn/submain.html#futuresys

when Future System finally imploded, the mad rush to get products back
into the 370 pipeline contributed to decision to pickup a lot of stuff I
had been doing (and was running widely inside the company) and release
it to customers.

Back in the 60s, various litigation contributed to the 23Jun1969
"unbundling" decision that started to charge for software, se services,
etc ... some past posts
http://www.garlic.com/~lynn/submain.html#unbundle

however, they managed to make the case that kernel software should still
be free. With the lack of products during the future system period and
clone processors getting market foothold ... a decision was made to
start charging for kernel software ... and the decision was made to make
my scheduler a separate kernel component and the guinea pig for starting
to charge for kernel software (as a result I got to spend a lot of time
with the business and legal people about policies for kernel software
charging).

later with the big explosion in online & interactive vm/4300 machines
... both with customers and internally ... the company made a decision
that vm/cms was the strategic interactive offering. It was then that the
TSO product manager asked me if I would port my dynamic adaptive
resource manager to MVS ... hoping that I could help fix the really
horrible TSO human factor characteristics ... old email reference:
http://www.garlic.com/~lynn/2006b.html#email800310

as I've mentioned periodically ... I declined the offer ... in part
because there was enormous number of other MVS issues (not just
scheduling) that affected its poor interactive characteristics.

as an aside, one of the problems I had (re)releasing my dynamic adaptive
resource manager ... was somebody from Armonk (with past history in POK
MVS) non-concurred with approval for the release because it didn't have
a lot of manual tuning knobs (because that was state-of-the-art at the
time with MVS having enormous number of manual tuning knobs). I tried to
explain that dynamic adaptive eliminated the necessity for all those
manual tuning knobs ... since it was repeatedly calculating them
dynamically adjusting for configuration and workload. I finally created
a "joke" ... I put in manual tuning knobs ... and described the
algorithms, code in detail as well as shipping source code. The "joke"
was that the dynamic adaptive code had more degrees of freedom than the
manual tuning knobs ... so any knob choice could be compensated for by
the dynamic code. All the code was also packaged in a source module I
named "STP" (after the television commercials about the "racer's edge").

-- 
virtualization experience starting Jan1968, online at home since Mar1970

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to